La résolution jusqu'à l'optimalité de problèmes d'optimisation combinatoire NP-difficiles nécessite une mise en oeuvre de méthodes de plus en plus complexes qui consomment de plus en plus de puissance de calcul. L'objectif de notre travail est de paralléliser un algorithme de "Branch and Cut" pour résoudre jusqu'à l'optimalité des instances difficiles du voyageur de commerce. Dans la première partie de notre travail, nous présentons les composantes principales de l'algorithme du "Branch and Cut". Nous étudions ensuite le problème du voyageur de commerce par une approche polyédrale. Nous donnons enfin une description détaillée de notre implémentation de l'algorithme du "Branch and Cut". Dans la deuxième partie, Nous commençons par une brève présentation du parallélisme, et un état de l'art des études menées sur la parallélisation de l'algorithme du "Branch and Bound". Puis, nous proposons plusieurs modèles de parallélisations de l'algorithme du "Branch and Cut". Nous décrivons ensuite la stratégie de contrôle de la recherche arborescente qu'on a adopté, les mécanismes de minimisation des coûts liés aux différentes étapes de la communication entre les processeurs et les stratégies d'équilibrages. Nous terminons en donnant les résultats obtenus sur le IBM-SP1.

We consider combinatorial t-design problems as discrete optimization problems. Our motivation is that only a few studies have been done on the use of exact optimization techniques in designs, and that classical methods in design theory have still left many open existence questions. Roughly defined, t-designs are pairs of discrete sets that are related following some strict properties of size, balance, and replication. These highly structured relationships provide optimal solutions to a variety of problems in computer science like error-correcting codes, secure communications, network interconnection, design of hardware; and are applicable to other areas like statistics, scheduling, games, among others. We give a new approach to combinatorial t-designs that is useful in constructing t-designs by polyhedral methods. The first contribution of our work is a new result of equivalence of t-design problems with a graph theory problem. This equivalence leads to a novel integer programming formulation for t-designs, which we call GDP. We analyze the polyhedral properties of GDP and conclude, among other results, the associated polyhedron dimension. We generate new classes of valid inequalities to aim at approximating this integer program by a linear program that has the same optimal solution. Some new classes of valid inequalities are generated as Chv´atal-Gomory cuts, other classes are generated by graph complements and combinatorial arguments, and others are generated by the use of incidence substructures in a t-design. In particular, we found a class of valid inequalities that we call stable-set class that represents an alternative graph equivalence for the problem of finding a t-design. We analyze and give results on the strength of these new classes of valid inequalities. We propose a separation problem and give its integer programming formulation as a maximum (or minimum) edge-weight biclique subgraph problem. We implement a pure cutting-plane algorithm using one of the stronger classes of valid inequalities derived. Several instances of t-designs were solved efficiently by this algorithm at the root node of the search tree. Also, we implement a branch-and-cut algorithm and solve several instances of 2-designs trying different base formulations. Computational results are included.

The minimum k-partition (MkP) problem is a well-known optimization problem encountered in various applications most notably in telecommunication and physics. Formulated in the early 1990s by Chopra and Rao, the MkP problem is the problem of partitioning the set of vertices of a graph into k disjoint subsets so as to minimize the total weight of the edges joining vertices in different partitions. In this thesis, we design and implement a branch-and-cut algorithm based on semidefinite programming (SBC) for the MkP problem. We describe and study the properties of two relaxations of the MkP problem, the linear programming and the semidefinite programming relaxations. We then derive a new strengthened relaxation based on semidefinite programming. This new relaxation provides tighter bounds compared to the other two discussed relaxations but suffers in term of computational time. We further devise an iterative clustering heuristic (ICH), a novel heuristic that finds feasible solution to the MkP problem and we compare it to the hyperplane rounding techniques of Goemans and Williamson and Frieze and Jerrum for k=2 and for k=3 respectively. Our computational results support the conclusion that ICH provides a better feasible solution for the MkP. Furthermore, unlike the hyperplane rounding, ICH remains very effective in the presence of negative edge weights. Next we describe in detail the design and implementation of a branch-and-cut algorithm based on semidefinite programming (SBC) to find optimal solution for the MkP problem. The ICH heuristic is used in our SBC algorithm to provide feasible solutions at each node of the branch-and-cut tree. Finally, we present computational results for the SBC algorithm on several classes of test instances with k=3, 5, and 7. Complete graphs with up to 60 vertices and sparse graphs with up to 100 vertices arising from a physics application were considered.

The minimum k-partition (MkP) problem is a well-known optimization problem encountered in various applications most notably in telecommunication and physics. Formulated in the early 1990s by Chopra and Rao, the MkP problem is the problem of partitioning the set of vertices of a graph into k disjoint subsets so as to minimize the total weight of the edges joining vertices in different partitions. In this thesis, we design and implement a branch-and-cut algorithm based on semidefinite programming (SBC) for the MkP problem. We describe and study the properties of two relaxations of the MkP problem, the linear programming and the semidefinite programming relaxations. We then derive a new strengthened relaxation based on semidefinite programming. This new relaxation provides tighter bounds compared to the other two discussed relaxations but suffers in term of computational time. We further devise an iterative clustering heuristic (ICH), a novel heuristic that finds feasible solution to the MkP problem and we compare it to the hyperplane rounding techniques of Goemans and Williamson and Frieze and Jerrum for k=2 and for k=3 respectively. Our computational results support the conclusion that ICH provides a better feasible solution for the MkP. Furthermore, unlike the hyperplane rounding, ICH remains very effective in the presence of negative edge weights. Next we describe in detail the design and implementation of a branch-and-cut algorithm based on semidefinite programming (SBC) to find optimal solution for the MkP problem. The ICH heuristic is used in our SBC algorithm to provide feasible solutions at each node of the branch-and-cut tree. Finally, we present computational results for the SBC algorithm on several classes of test instances with k=3, 5, and 7. Complete graphs with up to 60 vertices and sparse graphs with up to 100 vertices arising from a physics application were considered.

This thesis deals with the exact solution of large-scale minimum bisection problems via a semidefinite relaxation in a branch-and-cut framework. After reviewing known results on the underlying bisection cut polytope a study of new facet-defining inequalities is presented. They are derived from the known knapsack tree inequalities. We investigate strengthenings based on the new cluster weight polytope and present polynomial separation algorithms for special cases. The dual of the semidefinite relaxation of the minimum bisection problem is tackled in its equivalent form as an eigenvalue optimisation problem with the spectral bundle method. Implementational details regarding primal heuristics, branching rules, so-called support extensions for cutting planes and warm start are presented. We conclude with a computational study in which we show that our approach is competetive to state-of-the-art implementations using linear programming or semidefinite programming relaxations. / Diese Dissertation befasst sich mit der exakten Lösung großer Minimum Bisection Probleme über eine semidefinite Relaxierung in einem Branch-and-Cut Zugang. Nachdem bekannte Resultate zum zugrundeliegenden Bisection Cut Polytop dargestellt wurden, wird eine Studie neuer facettendefinierender Ungleichungen präsentiert. Diese werden von den bekannten Knapsack Tree Ungleichungen abgeleitet. Wir untersuchen Verstärkungen basierend auf dem neuen Cluster Weight Polytop und zeigen polynomiale Separierungsalgorithmen für Spezialfälle. Die Duale der semidefiniten Relaxierung des Minumum Bisection Problems wird in ihrer äquivalenten Form als Eigenwertoptimierungsproblem mit dem Spektralen Bündelverfahren bearbeitet. Details der Implementierung bezüglich primaler Heuristiken, Branchingregeln, sogenannter Supporterweiterungen für die Schnittebenen und Warmstart werden präsentiert. Wir beenden die Arbeit mit einer numerischen Studie, in der wir zeigen, dass unser Zugang konkurrenzfähig zu aktuellen Implementationen basierend auf linearen und semidefiniten Relaxierungen ist.

We consider combinatorial t-design problems as discrete optimization problems. Our motivation is that only a few studies have been done on the use of exact optimization techniques in designs, and that classical methods in design theory have still left many open existence questions. Roughly defined, t-designs are pairs of discrete sets that are related following some strict properties of size, balance, and replication. These highly structured relationships provide optimal solutions to a variety of problems in computer science like error-correcting codes, secure communications, network interconnection, design of hardware; and are applicable to other areas like statistics, scheduling, games, among others. We give a new approach to combinatorial t-designs that is useful in constructing t-designs by polyhedral methods. The first contribution of our work is a new result of equivalence of t-design problems with a graph theory problem. This equivalence leads to a novel integer programming formulation for t-designs, which we call GDP. We analyze the polyhedral properties of GDP and conclude, among other results, the associated polyhedron dimension. We generate new classes of valid inequalities to aim at approximating this integer program by a linear program that has the same optimal solution. Some new classes of valid inequalities are generated as Chv´atal-Gomory cuts, other classes are generated by graph complements and combinatorial arguments, and others are generated by the use of incidence substructures in a t-design. In particular, we found a class of valid inequalities that we call stable-set class that represents an alternative graph equivalence for the problem of finding a t-design. We analyze and give results on the strength of these new classes of valid inequalities. We propose a separation problem and give its integer programming formulation as a maximum (or minimum) edge-weight biclique subgraph problem. We implement a pure cutting-plane algorithm using one of the stronger classes of valid inequalities derived. Several instances of t-designs were solved efficiently by this algorithm at the root node of the search tree. Also, we implement a branch-and-cut algorithm and solve several instances of 2-designs trying different base formulations. Computational results are included.

University, Diss., 1999--Saarbrücken.

ADans cette thède, nous traitons le problème de stockage de conteneurs dans un terminal portuaire. Dans un premier temps, noux présentons une étude bibliographique dans laquelle sont analysés les travaux qui ont déhà été rélisé dans ce domaine. Ensuite, nous présentons une étude analytique, puis une modélisation mathématique et des méthodes de résolution numérique qui englobent des algorithmes efficaces. Nous proposons une démonstration de la compexité du problème de stockage de conteneurs en considérant différents cas de stockage. Ce problème étant "Np_difficile" peut être difficilement résolu avec le logiciel d'optimisation "ILOG CPLEX". », raison pour laquelle nous proposons un algorithme de banch-and-cut, qui est une méthode de résolution exacte et qui nous a permis de repousser les limites de "ILOG CPLEX". Nous avons aussi proposé des algorithmes métatheuristiques et des hybridations qui procurent des résultats très satisfaisants et qui sont très avantageux en temps de calcul. / AIn this thesis, we trait the container storage problem at port terminal. Initially, we present a state of the art in which the work that have been previously made in this filed are analyzed. After that, we present an analytical study. Thed we propose a mathematical modelling and some methods of resolution including effective algorithms. We propose a demonstration of the complexity of the problem by considering different cases of storage. This problme is "Np_difficult, so not always easy to solve by using the optimization software "ILOG CPLEX”. This is why we propose a branch-and-cut algorithm, wich is an optimal resolution algorithm and wich enables to go beyong the limits of "ILOG CPLEX". We also proposed meta-heuristic algorithms and hybridizations wich provide satisfactory resulys and wich required less calculation times.

Orientador: Vinicius Amaral Armentano / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Eletrica e de Computação / Made available in DSpace on 2018-08-10T01:09:54Z (GMT). No. of bitstreams: 1 Fonseca_SaraLuisadeAndrade_M.pdf: 1363535 bytes, checksum: aa7c01c779a21ea25aa3b603425c92fe (MD5) Previous issue date: 2007 / Resumo: Em 1958, Gomory propôs uma desigualdade válida ou corte a partir do tableau do método simplex para programação linear, que foi utilizado no primeiro método genérico para resolução de problemas de programação inteira. Em 1960, o corte foi estendido para problemas de programação inteira mista. Em 1973, Chvátal sugeriu um corte derivado da formulação original do problema de programação inteira, e devido à equivalência com o corte de Gomory, este passou a ser chamado de corte de Chvátal-Gomory. A importância do corte de Gomory só foi reconhecida em 1996 dentro do contexto do método branch-and-cut para resolução de problemas de programação inteira e programação inteira mista. Desde então, este corte é utilizado em resolvedores comerciais de otimização. Recentemente, diversos cortes novos derivados do corte de Chvátal-Gomory foram propostos na literatura para programação inteira. Este trabalho trata do desenvolvimento de algoritmos para alguns destes cortes, e implementação computacional em um contexto de branch-and-cut, no ambiente do resolvedor CPLEX. A eficácia dos cortes é testada em instâncias dos problemas da mochila multidimensional, designação generalizada e da biblioteca MIPLIB. / Abstract: In 1958, Gomory proposed a valid inequality or cut from the tableau of the simplex method for linear programming, which was used in the first generic method for solving integer programming problems. In 1960, the cut was extended to handle mixed integer programming problems. In 1973, Chvátal suggested a cut that is generated from the original formulation of an integer programming problem, and due to the equivalence with the Gomory cut, it was named Chvátal-Gomory cut. The importance of the Gomory cut was recognized only in 1996 in the context of the branch-and-cut method for solving (mixed) integer programming problems. Today, such a cut is utilized in optimization commercial solvers. Recently, several new cuts derived from the Chvátal-Gomory cut have been proposed in the literature for integer programming. This work deals with the development of algorithms and computational implementations for some of the new proposed cuts, in a context of the branch-and-cut method, by using the CPLEX solver. The efficiency of the cuts is tested on instances of the multi-dimensional knapsack, generalized assignment problems, and instances from the MIPLIB library. / Mestrado / Automação / Mestre em Engenharia Elétrica

This dissertation studies in depth the structure of the Discrete Ordered Median Problem (DOMP), to define new formulations and resolution algorithms. Furthermore we analyze an interesting extension for DOMP, namely MDOMP (Monotone Discrete Ordered Median Problem). This thesis is structured in three main parts.First, a widely theoretical and computational study is reported. It presents several new formulations for the Discrete Ordered Median Problem (DOMP) based on its similarity with some scheduling problems. Some of the new formulations present a considerably smaller number of constraints to define the problem with respect to some previously known formulations. Furthermore, the lower bounds provided by their linear relaxations improve the ones obtained with previous formulations in the literature even when strengthening is not applied. We also present a polyhedral study of the assignment polytope of our tightest formulation showing its proximity to the convex hull of the integer solutions of the problem. Several resolution approaches, among which we mention a branch and cut algorithm, are compared. Extensive computational results on two families of instances, namely randomly generated and from Beasley's OR-library, show the power of our methods for solving DOMP. One of the achievements of the new formulation consists in its tighter LP-bound. Secondly, DOMP is addressed with a new set partitioning formulation using an exponential number of variables. This chapter develops a new formulation in which each variable corresponds to a set of demand points allocated to the same facility with the information of the sorting position of their corresponding distances. We use a column generation approach to solve the continuous relaxation of this model. Then, we apply a branch-cut-and-price algorithm to solve to optimality small to moderate size of DOMP in competitive computational time.To finish, the third contribution of this dissertation is to analyze and compare formulations for the monotone discrete ordered median problem. These formulations combine different ways to represent ordered weighted averages of elements by using linear programs together with the p-median polytope. This approach gives rise to two efficient formulations for DOMP under a hypothesis of monotonicity in the lambda vectors. These formulations are theoretically compared and also compared with some other formulations valid for the case of general lambda vector. In addition, it is also developed another new formulation, for the general case, that exploits the efficiency of the rationale of monotonicity. This representation allows to solve very efficiently some DOMP instances where the monotonicity is only slightly lost. Detailed computational tests on all these formulations is reported in the dissertation. They show that specialized formulations allow to solve to optimality instances with sizes that are far beyond the limits of those that can solve in the general case. / Cette dissertation étudie en profondeur la structure du "Discrete Ordered Median Problem" (DOMP), afin de proposer de nouvelles formulations et de nouveaux algorithmes de résolution. De plus, une extension intéressante du DOMP nommée MDOMP ("Monotone Discrete Ordered Median Problem") a été étudiée.Cette thèse a été structurée en trois grandes parties.La première partie présente une étude riche aux niveaux théorique et expérimentale. Elle développe plusieurs formulations pour le DOMP qui sont basées sur des problèmes d'ordonnancement largement étudiés dans la littérature. Plusieurs d'entres elles nécessitent un nombre réduit de contraintes pour définir le problème en ce qui concerne certaines formulations connues antérieurement. Les bornes inférieures, qui sont obtenues par la résolution de la relaxation linéaire, donnent de meilleurs résultats que les formulations précédentes et ceci même avec tout processus de renforcement désactivé. S'ensuit une étude du polyhèdre de notre formulation la plus forte qui montre sa proximité entre l'enveloppe convexe des solutions entières de notre problème. Un algorithme de branch and cut et d'autres méthodes de résolution sont ensuite comparés. Les expérimentations qui montrent la puissance de nos méthodes s'appuient sur deux grandes familles d'instances. Les premières sont générées aléatoirement et les secondes proviennent de Beasley's OR-library. Ces expérimentations mettent en valeur la qualité de la borne obtenue par notre formulation.La seconde partie propose une formulation "set partitioning" avec un nombre exponentiel de variables. Dans ce chapitre, la formulation comporte des variables associées à un ensemble de demandes affectées à la même facilité selon l'ordre établi sur leurs distances correspondantes. Nous avons alors développé un algorithme de génération de colonnes pour la résolution de la relaxation continue de notre modèle mathématique. Cet algorithme est ensuite déployé au sein d'un Branch-and-Cut-and-Price afin de résoudre des instances de petites et moyennes tailles avec des temps compétitifs.La troisième partie présente l'analyse et la comparaison des différentes formulations du problème DOMP Monotone. Ces formulations combinent plusieurs manières de formuler l'ordre des éléments selon les moyennes pondérées en utilisant plusieurs programmes linéaires du polytope du p-median. Cette approche donne lieu à deux formulations performantes du DOMP sous l'hypothèse de monotonie des vecteurs lambda. Ces formulations sont comparées de manière théorique puis comparées à d'autres formulations valides pour le cas général du vecteur lambda. Une autre formulation est également proposée, elle exploite l'efficacité du caractère rationnel de la monotonie. Cette dernière permet de résoudre efficacement quelques instances où la monotonie a légèrement disparue. Ces formulations ont fait l'objet de plusieurs expérimentations dècrites dans ce manuscrit de thèse. Elles montrent que les formulations spécifiques permettent de résoudre des instances plus importantes que pour le cas général. / Este trabajo estudia en profundidad la estructura del problema disctreto de la mediana ordenada (DOMP, por su acrónimo en inglés) con el objetivo de definir nuevas formulaciones y algoritmos de resolución. Además, analizamos una interesante extensión del DOMP conocida como el problema monótono discreto de la mediana ordenada (MDOMP, de su acrónimo en inglés).Esta tesis se compone de tres grandes bloques.En primer lugar, se desarrolla un detallado estudio teórico y computacional. Se presentan varias formulaciones nuevas para el problema discreto de la mediana ordenada (DOMP) basadas en su similaridad con algunos problemas de secuenciación. Algunas de estas formulaciones requieren de un cosiderable menor número de restricciones para definir el problema respecto a algunas de las formulaciones previamente conocidas. Además, las cotas inferiores proporcionadas por las relajaciones lineales mejoran a las obtenidas con formulaciones previas de la literatura incluso sin reforzar la nueva formulación. También presentamos un estudio poliédrico del politopo de asignación de nuestra formulación más compacta mostrando su proximidad con la envolvente convexa de las soluciones enteras del problema. Se comparan algunos procedimientos de resolución, entre los que destacamos un algoritmo de ramificación y corte. Amplios resultados computacionales sobre dos familias de instancias -aleatoriamente generadas y utilizando la Beasley's OR-library- muestran la potencia de nuestros métodos para resolver el DOMP.En el segundo bloque, el problema discreto de la mediana ordenada es abordado con una formulación de particiones de conjuntos empleando un número exponencial de variables. Este capítulo desarrolla una nueva formulación en la que cada variable corresponde a un conjunto de puntos de demanda asignados al mismo servidor con la información de la posición obtenida de ordenar las distancias correspondientes. Utilizamos generación de columnas para resolver la relajación continua del modelo. Después, empleamos un algoritmo de ramificación, acotación y "pricing" para resolver a optimalidad tamaños moderados del DOMP en un tiempo computacional competitivo.Por último, el tercer bloque de este trabajo se dedica a analizar y comparar formulaciones para el problema monótono discreto de la mediana ordenada. Estas formulaciones combinan diferentes maneras de representar medidas de pesos ordenados de elementos utilizando programación lineal junto con el politopo de la $p$-mediana. Este enfoque da lugar a dos formulaciones eficientes para el DOMP bajo la hipótesis de monotonía en su vector $lambda$. Se comparan teóricamente las formulaciones entre sí y frente a algunas de las formulaciones válidas para el caso general. Adicionalmente, se desarrolla otra formulación válida para el caso general que explota la eficiencia de las ideas de la monotonicidad. Esta representación permite resolver eficientemente algunos ejemplos donde la monotonía se pierde ligeramente. Finalmente, llevamos a cabo un detallado estudio computacional, en el que se aprecia que las formulaciones ad hoc permiten resolver a optimalidad ejemplos cuyo tamaño supera los límites marcados en al caso general. / Doctorat en Sciences / info:eu-repo/semantics/nonPublished